The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, hardcoded credentials, or data exfiltration attempts were found across the skill files.
[INDIRECT_PROMPT_INJECTION]: A surface for indirect prompt injection exists in the scripts/new_design.py script, which incorporates user-provided inputs into generated markdown documents without sanitization.
Ingestion points: User inputs for system name and features in scripts/new_design.py and through the agent interaction.
Boundary markers: The generated output lacks explicit delimiters or warnings to ignore instructions within the user-provided content.
Capability inventory: The script has the capability to write files to the local system using pathlib (scripts/new_design.py).
Sanitization: There is no evidence of input validation or escaping for the user-provided strings before they are interpolated into the document template.
[COMMAND_EXECUTION]: The provided Python script scripts/new_design.py is used to generate and save design documents locally. It performs file writing based on user-specified paths, which is expected functionality for a utility script but is noted as a capability that interacts with the filesystem.

system-design-interview